Today dominant technology for detection
and classification of products based upon appearance
feature is the technology of image processing.
Generally this operation is performed in two main
phases, namely feature extraction and classification.
In the former by determining the required features and
by choosing parameters and method of extraction, the
features are detected from a raw image and then
optimized. In the latter phase, the regions with similar
tissues are determined and the border of different
tissues is detected. The main goal of feature extraction
of tissues is to provide a criterion for detecting tissue
properties of image, such as softness and roughness,
cognation, flatness, frequency, being in order and so
forth. There are different methods for extraction of
tissue features; each one has its pros and cons with
regard to its application, speed, accuracy, etc.
Techniques like concurrent matrices or self –
correlation function and approaches based on model
are not applicable, because of complexity of
calculation time for real – time inspection systems.
Nevertheless, in most studies, these are used as
methods with high accuracy. In this paper; the method
of signal processing is utilized. The application is
possible in two ways:
Today dominant technology for detection
and classification of products based upon appearance
feature is the technology of image processing.
Generally this operation is performed in two main
phases, namely feature extraction and classification.
In the former by determining the required features and
by choosing parameters and method of extraction, the
features are detected from a raw image and then
optimized. In the latter phase, the regions with similar
tissues are determined and the border of different
tissues is detected. The main goal of feature extraction
of tissues is to provide a criterion for detecting tissue
properties of image, such as softness and roughness,
cognation, flatness, frequency, being in order and so
forth. There are different methods for extraction of
tissue features; each one has its pros and cons with
regard to its application, speed, accuracy, etc.
Techniques like concurrent matrices or self –
correlation function and approaches based on model
are not applicable, because of complexity of
calculation time for real – time inspection systems.
Nevertheless, in most studies, these are used as
methods with high accuracy. In this paper; the method
of signal processing is utilized. The application is
possible in two ways:
การแปล กรุณารอสักครู่..